Reef manta rays (Mobula alfredi), are declining worldwide, in large part due to fishing pressure for their gill rakers [1,2]. Despite significant advances in our knowledge and understanding of this species in the past decade [3], more detailed information on the biology and the ecology of this species throughout its range is urgently needed [4]. Specifically, data on spatiotemporal dynamics and habitat use are necessary to develop concrete management plans and conservation actions [4] to prevent further declines of reef manta rays, now listed as “vulnerable” on the IUCN Red-List [5]. Satellite telemetry using pop-up satellite archival tags (PSAT tags) is one of the most effective methods to investigate fine scale horizontal and vertical movements and habitat use in manta rays [4,6,7–9], but until now there have been no such studies conducted in New Caledonian waters.

As planktivores, manta rays spend a major part of their time feeding or searching for foraging grounds [3,10–13]. Manta ray aggregations have been observed and monitored in multiple locations in tropical and sub-tropical waters around the world [3,12,14,16,17]. Seasonal or long-term presence of the species on a particular site is often associated with enhanced local productivity and increased food availability. For instance, seasonal migrations were found to be correlated with monsoonal shifts in the Indian Ocean [12,17]. As opportunistic feeders, manta rays are capable of undertaking relatively large-scale movements between productive areas (up to 750 km) [3,4,8,13–17]. Some studies have shown that reef manta rays are also able to explore substantial depths (up to 432 m), presumably to feed on deeper zooplankton and other food resources [8,10,18–20]. These foraging strategies remain unclear and more detailed information on this behaviour and the associated drivers are needed.

In New Caledonia, reef manta rays are not targeted by fishing, but have a highly fragmented distribution due to the specificity of their food resources and preferred habitat [8,13,16,21,22]. Environmental processes and conditions shape the distribution and the abundance of their zooplankton prey [21–24]. Nutrient enrichment is known to be the primary factor of phytoplankton proliferation, causing a subsequent increase in zooplankton abundance. Eutrophication benefit the development of phytoplankton upon which zooplankton feed [25,26]. Processes such as coastal upwellings and river run-off are both important sources of nutrient enrichment of coastal waters [26–28]. These processes, combined with tidal currents and bathymetry can support dense zooplankton concentrations and favourable feeding grounds for filter feeders such as Mobula alfredi [20]. Massive feeding aggregation of hundreds of reef mantas have been observed targeting such dense zooplankton aggregations in the Maldives [29,30] and occasionally in the southern reaches of the Great Barrier Reef [22]. In New Caledonia, manta feeding grounds seem to be scattered, with aggregations never exceeding a dozen individuals (Lassauce, pers. obs.).

This short communication presents the first data collected on the diving behaviour of reef manta rays in New Caledonia. These data reveal an unexpected outstanding feature: the unique depths and high number of deep dives, which considerably extend the known depth range for Mobula alfredi.

Of the 11 tags deployed, two (#167756 and #162378) failed to transmit to the Argos system. The deployment duration of the functioning tags (n = 9) ranged from 3 to 174 days (73 ± 58 days). On average, 78 ± 19% of the data recorded by the tags was either transmitted by ARGOS satellite or downloaded from two tags recovered after deployment (Table 1). All nine individuals recorded dives deeper than 300 m (n = 78), and six of them performed dives deeper than 450 m (n = 22), including two exceptionally deep dives by two of the smaller tagged individuals (2.4 m female CD-MA-0167 and 3 m male CD-MA-0109) that reached maximum depths of 624 ± 4 m and 672 ± 4 m, respectively (Fig 2). This last dive extends the reported depth range for M. alfredi by more than 200 m, previously recorded as 432 m in the Red Sea [8]. A similar study in Indonesia using the same tags and tagging technique recorded a reef manta ray reaching a maximum depth of 624 ± 4 m in East Kalimantan (Erdmann, unpub.). In this study only 6 of the 30 tagged manta rays recorded dives deeper than 300 m, which indicate fewer deep dives compared to New Caledonia. In the Red Sea, 5 of the 7 tagged individuals dived deeper than 300 m [8] and none of the tracked manta rays in Eastern Australia reached these depths [7]. In New Caledonia, all individuals dived deeper than 300 m, representing 7.1% of all dives (n = 1099). The 200 m level was reached in 13% of all dives (n = 1099). The mean depth of all the dives was 103.1 ± 104.9 m (n = 1099) (Table 2). Maximum depths recorded were not correlated with deployment duration (Pearson’s r test, r = -0.19, n = 9, p > 0.5). The dives recorded by the New Caledonian mantas are thus both deeper in an absolute sense and more frequently exceeding the 200m mark than previously found in other parts of the world [7,8,32].

Difference in mean depths per individual between day and night were only significant for three manta rays. Regarding the diel comparison of the overall distribution of the number of dives per depth range, no significant difference (p > 0.05) was observed (Table 2).

Among all individuals, the number of deep dives (depth > 300 m) was only significantly larger at night (7 pm– 7 am) than during the day (7 am– 7 pm) (Fisher exact test, p = 0.010) (Fig 2). This behaviour could be explained by the nocturnal exploitation, at night, of demersal food sources, which has been observed in reef manta rays [4,18,19], oceanic manta rays [33], other mobulid species [34–36], as well as whale sharks [37]. In this study, all manta rays spent a relatively short amount of time at maximum depth during their absolute deepest dive. Bottom time during each manta’s absolute deepest dive averaged 11 ± 7 minutes, varying from 26.6 minutes at 376 ± 4 m to 1.4 minutes at 304 ± 4 m (Table 2). There was no significant correlation between maximum depth reached and the time spent at this depth (Pearson’s r test, r = - 0.06, n = 9, p > 0.5).

Analysis of dive profiles can provide valuable information on diving behaviour [38]. Classification of dive profiles has been mainly conducted on air-breathing marine animals such as seabirds [39], sea turtles [40], or seals [41], as well as a few studies focused on predatory fish [42,43]. These analyses revealed two main patterns that have been associated with distinct behaviours. Dives with very short or no bottom time, called “V-Shaped” dives, are possible indicators of travelling and/or prey searching behaviour [38–42]. By comparison, U-shaped or square-shaped dives profiles with distinctively longer bottom times are thought to suggest foraging activities [38–42]. Asymmetrical V-shaped dives were described for reef manta rays by Braun et al. [8]. These authors suggested that short bottom times with relatively slow descents and faster ascents reflected an optimized travelling behaviour using gliding [8]. In this study, three manta rays showed this type of profile with very limited time spent at maximum depth during their deepest dive (1.4 min at 304 ± 4 m, 2.2 min at 672 ± 4 m and 6.5 min at 384 ± 4 m) (Table 2). While travelling and/or prey searching could be an explanation for these particular dives, additional data on the velocity during ascent and descent would be needed to test this hypothesis. On the other hand, our results show that six manta rays remained at maximum depths for more than 10 minutes. These dive patterns are more akin to U-shaped profiles, suggesting the exploitation of aggregated prey [38,42]. As fishes, manta rays diving is not limited by the ability to store oxygen, but more probably by the low temperatures at these depths. During dives, temperatures were always colder than 20°C below 300 m, with a minimum temperature of 7.6°C recorded at the maximum depth of 672 ± 4 m (Fig 3). Manta rays are poikilothermic species with an optimal thermal range from 20 to 26°C [3,7,32]. Previous studies have also shown that mobulid rays have the capacity to transmit warmth to the brain using a specific vesicular network in the pectoral fin that can function as a counter‐current heat exchanger [44]. Consequently, basking in warm shallow water prior to diving and active swimming during descent and ascent could be used to increase the body temperature. This mechanism would allow manta rays to produce enough heat to reach demersal food resources and feed for a relatively short amount of time despite the cold temperatures of these depths. This behaviour has been observed for other mobulid rays and other fish such as tunas and sharks [33–37,44–47]. In order to fully support this hypothesis, more detailed data on the dive profile of these manta rays are necessary to confirm rapid descent and slower ascent directly followed by an extended period of basking in warm shallow water. If this last hypothesis can be verified, the identification of such comportment for the reef manta rays of New Caledonia highlights the probable presence of important demersal food resources at depth, resulting in significant foraging success that presumably compensates for the energetic costs. The unusual depths reached and number of deep dives recorded suggest that foraging opportunities could be insufficient in the upper layer of the water column in New Caledonian waters, thereby forcing manta rays to explore deeper food resources. Detailed data on resource availability at varying depths and on the diet of manta rays in this region will help in determining the underlying drivers of their movements.

These preliminary results extend the global knowledge on the depth range and, more generally, the habitat use of M. alfredi. In this case, these data appear to support previous findings that prey at mesopelagic depths (from 200 m to 1000 m) [48] are valuable, if not indispensable, food resources for reef manta rays [10,18,19]. A comprehensive knowledge of the distribution and the habitat use of the reef mantas is necessary to inform conservation and fisheries management measures to ensure the long-term survival of the species [4]. Protective legislation has improved in recent years and numerous marine protected areas (MPAs) have been created throughout the known range of reef manta rays [49]; however, many of these MPAs are coastal in nature and do not extend into the deeper offshore waters used by reef mantas. As deepwater fisheries are increasingly exploiting the mesopelagic zone [50], our study highlights the importance of incorporating offshore waters and deep-water foraging grounds in manta conservation initiatives.